Online Trading Competition 2018

Transcription

1 Online Trading Competition 2018 Table of Contents December 2 nd, 2018 Case Package

2 Table of Contents Table of Contents... 2 Important Information... 3 Case Summaries... 4 Sales & Trader Case... 5 Volatility Trading Case Algorithmic Trading Case

3 Important Information Practice Servers Practice servers will be made available starting on November 9 th, 2018 and will operate 24 hours a day 7 days a week until the start of the competition. Information on how to download and install the RIT v2.0 Client is available on the RIT website. We will post information on how to login to any server port on the ROTC website. Remember that you can type in any username and password and it will automatically create an account if it does not exist. If you have forgotten your password or the username appears to be taken, simply choose a new username and password to create a new account. Important Information Scoring and Ranking Methodology The Scoring and Ranking Methodology document will be released prior to the start of the competition on the ROTC website. Schedule and Timeline The competition is held online, it will start at 10:00 am on Sunday, December 2 nd and end at 1:00 pm and it will be run until the number of heats described further in this document is completed. The ROTC organizers will not wait for anyone who is late. Competition servers will be set up 10 minutes before the start. In the event of a server failure, the case session will be rerun and the scores from that session will not be kept. 3

4 Case Summaries Sales and Trader Case The Sales and Trader Case challenges participants to put their critical thinking and analytical abilities to the test in an environment that requires them to evaluate the liquidity risk associated with different tender offers. Participants will be faced with multiple tender offers throughout the case. This will require participants to make rapid judgments on the profitability and subsequent execution, or rejection, of each offer. Profits can be generated by taking advantage of price differentials between market prices and prices offered in the tender offers. Once any tender has been accepted, participants should aim to efficiently close out the large positions to maximize returns. Case Summaries Volatility Trading Case The Volatility Trading Case will test participants on their ability to implement volatility trading strategies through options. This will be achieved through positions in mispriced options whilst hedging delta exposure. Taking place over a one month simulation period, participants will receive accurate estimates on forecasted volatility for underlying assets. Participants will need to analyze market information and update expectations in order to execute a profitable strategy. Algorithmic Trading Case The Algorithmic Trading Case is designed to challenge participants programming skills since they are required to develop algorithms using RIT API to automate the market-making process and react to changing market conditions. Throughout the case, these algorithms will submit orders to profit from market-making and also from any arbitrage opportunities that may arise. Due to the highfrequency nature of the case, participants will have to develop algorithms to adapt to changes in market dynamics. 4

5 Sales & Trader Case Overview The Sales and Trader Case challenges participants to put their critical thinking and analytical abilities to the test in an environment that requires them to evaluate the liquidity risk associated with different tender offers. Participants will be faced with multiple tender offers throughout the case. This will require participants to make rapid judgments on the profitability and subsequent execution, or rejection, of each offer. Profits can be generated by taking advantage of price differentials between market prices and prices offered in the tender offers. Once any tender has been accepted, participants should aim to efficiently close out the large positions to maximize returns. Sales & Trader Case Description The trading session will consist of five, 10-minute heats with each heat to be independently traded and representing one month of calendar time. Each heat will have a unique objective and could involve up to four securities with different volatility and liquidity characteristics. Parameter Value Number of trading heats 5 Trading time per heat 600 seconds (10 minutes) Calendar time per heat 1 month (20 trading days) Tender offers will be generated by computerized traders and distributed at random intervals to random participants. Participants must subsequently evaluate the profitability of these tenders when accepting or bidding on them. Order submission using the RIT API will be disabled. Only data retrieval via Real-Time Data (RTD) links and the RIT API will be enabled. Market Dynamics There are five heats, each with unique market dynamics and parameters. Potential parameter changes include factors such as spread of tender orders, liquidity, and volatility. Market dynamics and parameter details regarding each heat can be found below, allowing participants to formulate trading strategies. 5

6 Heat 1 Stock Price Commissions Tender Spread Volatility Liquidity SUIT $30 $0.02 High Low Medium TIE $15 $0.01 Medium Medium Medium Heat 2 Stock Price Commissions Tender Spread Volatility Liquidity RED $15 $0.02 High Low High BLUE $30 $0.02 Medium Medium Medium Heat 3 Stock Price Commissions Tender Spread Volatility Liquidity NYC $10 $0.02 Medium Medium Medium LNDN $15 $0.02 Medium Medium Medium PRS $25 $0.01 High High Medium Sales & Trader Case Heat 4 Stock Price Commissions Tender Spread Volatility Liquidity DWHT $25 $0.02 High High Medium MCHL $30 $0.02 Medium High High CRED $40 $0.03 Medium Medium High Heat 5 Stock Price Commissions Tender Spread Volatility Liquidity VNLLA $55 $0.03 Medium Medium Medium CHOCO $45 $0.02 Medium High Low ICRM(ETF) $100 $0.05 High High Medium Note: Heat 5 includes a third security which is an ETF. The ETF ICRM is comprised equally of securities VNLLA and CHOCO. This means that at any given point in time the price of ICRM should be the combined price of VNLLA and CHOCO. The implication of this ETF also means that ICRM will close out at the price equal to the price of VNLLA and CHOCO if held until the end of the case instead of the last traded price. During each sub-heat, participants will occasionally receive one of three different types of tender offers: private tenders, competitive auctions, and winner-take-all tenders. Tender offers are generated by the server and randomly distributed to random participants at different times. Each participant will get the same number of tender offers with variations in price and quantity. No trading commission will be paid on tenders. Private Tenders are routed to individual participants and are offers to purchase or sell a fixed volume of stock at a fixed price. The tender price is influenced by the current market price. 6

7 Competitive Auction offers will be sent to every participant at the same time. Participants will be required to determine a competitive, yet profitable price to submit for a given volume of stock from the auction. Any participant that submits an order that is better than the base-line reserve price (hidden from traders) will automatically have their order filled, regardless of other participants bids. If accepted, the fills will occur at the price that the participant submits. Winner-take-all Tenders request participants to submit bids to buy or sell a fixed volume of stock. After all prices have been received, the tender is awarded to the single highest bidder or lowest offer. The winning price however must meet a base-line reserve price. If no offer meets the reserve price, then the trade may not be awarded to anyone (i.e. if all participants bid $2.00 for a $10 stock, nobody will win). Sales & Trader Case Calculation of the Profit or Loss of Traders The prices generated by the RIT for this case follow a random walk process using a return drawn from a normal distribution with a mean of zero. That is, at any point in the simulation, the probability that the price will go up is equal to the probability that the price will go down. This means that participants cannot predict the future price of the stocks without taking a bet. Therefore, the ROTC scoring committee will consider buying (selling) stocks for reasons other than reducing the exposure associated with accepting a tender offer to be equivalent to speculating (taking a bet) on the price movement. These types of trades will be marked as speculative trades. Participants will have time to think about the offer before they accept it. For example, one may receive a tender offer at time tt = 0 and will have until tt = 30 to decide whether to accept. Any trades made by a participant during this time without accepting the tender offer will be considered as front-running 1 since the participant had the advance knowledge of a pending institutional order. The ROTC scoring committee will mark these trades as front-running trades. This case is designed to only reward the participants for identifying, accepting, and closing out 2 tender offer positions at a profit while managing liquidity risk and execution risk. Any other strategy will not be considered. In particular, the total profit of each participant 3 will be categorized into two 1 Front-running is the unethical and illegal practice of trading a security for your own account while taking advantage of the information contained in the pending orders from your institutional clients. 2 Closing out a position means that a participant is executing a trade that is the opposite of the current position in order to eliminate the exposure. 3 Total profit of each participant is the profit (or loss) that you can observe in the RIT at the end of each heat. 7

8 parts: profits from tender offers and profit from speculation ; the latter category includes the profits that are a result of either speculative trades or front-running trades. Profits from tenders are the profits (or losses) gained from efficiently closing out the position from accepted tenders into the market. Profits from speculation are profits (or losses) generated through trades that are not associated with tenders (speculative trades or front-running trades). An adjusted P&L will be calculated based on the following formula: AAAAAAAAAAAAAAAA PP&LL = PPPPPPPPPPPP FFFFFFmm TTTTTTTTTTTTTT + MMMMMM(0, PPPPPPPPPPPP FFFFFFFF SSSSSSSSSSSSSSSSSSSSSS) Sales & Trader Case Participants will be ranked and scored based on their AAAAAAAAAAAAAAAA PP&LL. For example, consider a participant who has made $10,000 from tenders and $50,000 from speculation, the total profit is $60,000 (= $10,000 + $50,000) but the AAAAAAAAAAAAAAAA PP&LL will be only $10,000 [= $10,000 + mmmmmm(0, $50,000)]. Another example, consider a participant who has made $35,000 from tenders and lost $20,000 from speculation (PPPPPPPPPPPP FFrrrrrr SSSSSSSSSSSSSSSSSSSSSS = $20,000) ; the total profit is $15,000 ($35,000 $20,000) and it is the same as the AAAAAAAAAAAAAAAA PP&LL [$15,000 = $35,000 + mmmmmm(0, $20,000)]. From the last example, please note that any losses from speculation will still be considered and therefore, negatively affect your AAAAAAAAAAAAAAAA PP&LL. The AAAAAAAAAAAAAAAA PP&LL will be calculated by the ROTC scoring committee at the end of each heat and it will not be included in the P&L calculation in RIT. However, participants will be provided with an Excel tool 4, the Performance Evaluation Tool, that will allow them to calculate the AAAAAAAAAAAAAAAA PP&LL. Trading Limits and Transaction Costs Each participant will be subject to gross and net trading limits: the net and gross trading limits for all of the versions are NET 250,ooo shares, or GROSS 250,000 shares. The gross trading limit reflects the sum of the absolute values of the long and short positions across all securities, while the net trading limit reflects the sum of long and short positions such that short positions negate any long positions. Trading limits will be strictly enforced and participants will not be able to exceed them. The maximum trade size will be 25,000 shares, restricting the volume of shares transacted per trade to 25, The Performance Evaluation Tool is uploaded on the ROTC website. 8

9 There is a maximum stop loss of $1.5 Million per person for each trading heat. If a participant loses more than $1.5 Million, he/she will be forced to stop trading for the remainder of the heat. Position Close-Out Any non-zero position will be closed out at the end of trading based on the last traded price. This includes any long or short position open in any security. Computerized market makers will increase the liquidity in the market towards the end of trading to ensure the closing price cannot be manipulated. Sales & Trader Case Key Objective Evaluate the profitability of tender offers by analyzing the market liquidity. Participants will accept the tenders that will generate positive profits while rejecting the others. Submit competitive, yet profitable, bids and offers on above reserve and winner-take-all tenders to maximize potential profits while managing liquidity and market risk. There is a chance that the market may move away from your transactions prices, so maintaining large short or long positions may result in losses. Use a combination of limit orders, market orders and marketable limit orders to mitigate any liquidity and price risks from holding open positions. 9

10 Volatility Trading Case Overview The Volatility Trading case will test competitors on their ability to implement volatility trading strategies through options. This will be achieved through positions in mispriced options whilst hedging delta exposure. Taking place over a one month simulation period, competitors will receive accurate estimates on forecasted volatility for underlying assets. Participants will need to analyze market information and update expectations in order to execute a profitable strategy. Description The trading session will consist of five, 10-minute heats with each heat to be independently traded and representing one month of calendar time. Each heat will have different volatility characteristics which participants must subsequently evaluate to identify mispricing. Volatility Trading Case Case Parameter Value Number of trading heats 5 Trading time per heat Calendar time per heat 600 seconds (10 minutes) 1 month (20 trading days) Order submission using the RIT API will be disabled. Only data retrieval via Real-Time Data (RTD) links and the RIT API will be enabled. Market Dynamics Participants will be able to trade RTM and 10 separate options contracts on RTM. All options are European, so early exercise is not allowed. Call Ticker Strike Price Put Ticker RTM48C 48 RTM48P RTM49C 49 RTM49P RTM50C 50 RTM50P RTM51C 51 RTM51P RTM52C 52 RTM52P All securities are priced by a very large market-maker who will always quote a bid-ask spread of 2 cents (i.e. $49.99*$50.01 for the RTM, or $4.08*$4.10 for the RTM46C). The bids and asks are for an infinite quantity (there are no liquidity constraints in this case). 10

11 The price of RTM is a random-walk and the path is generated using the following process: PP RRRRRR,tt = PP RRRRRR,tt 1 (1 + rr tt ) wwheeeeee rr tt ~NN(0, σσ tt ) The price of the stock is based on the previous price multiplied by a return that is drawn from a normal distribution with a mean of zero and standard deviation (volatility) of σσ tt = 20% (on an annualized basis). The trading period is divided into 4 weeks, with tt = being week one, tt = being week two, and so on. At the beginning of each week, the volatility value (σσ tt ) will shift, and the new value will be provided to participants. In addition, at the middle of each week (e.g. tt = 75) an analyst estimate of next week s volatility value will be announced. Volatility Trading Case The observed and tradable prices of the options will be based on a computerized market-maker posting bids and offers for all options. The market maker will price the options using the Black-Scholes model. It is important to note that the case assumes a risk-free rate of 0%. The volatility forecasts made by the market maker are uninformed and therefore will not always accurately reflect the future volatility of RTM. Mispricing will occur, creating trading opportunities for market participants. These opportunities could be between specific options with respect to other options, specific options with respect to the underlying, or all options with respect to the underlying. The focus of this case is on trading volatility without being exposed to price changes of the underlying security, RTM. Participants are therefore asked to manage their portfolio s delta exposure. Recognizing the transactions costs and impracticality of perfect delta hedging (i.e. keeping the portfolio s delta at zero at all times), the ROTC scoring committee will allow the portfolio s delta to be different from zero but it requires it to stay between dddddddddd llllllllll and + dddddddddd llllllllll. Please note that dddddddddd llllllllll is an integer number greater than 1,000 that will be announced at the beginning of the case via a news release in the RIT. For example, the following news could be released: The delta limit for this sub-heat is 5,000. According to that news, any participant that has a portfolio s delta greater than 5,000 will be penalized according to the penalties explained below. For every second that a participant exceeds the limit (+/ dddddddddd llllllllll), s/he will be charged a penalty according to the following formula: PPPPPPPPPPPPPP aaaa ssssssssssss tt = ΔΔ pp,tt dddddddddd llllllllll iiii ΔΔ pp,tt > dddddddddd llllllllll 0 iiii ΔΔ pp,tt dddddddddd llllllllll Where, ΔΔ pp,tt is the portfolio s delta at time tt Penalties will be applied at the end of each heat and will not be included in the P&L calculation in RIT. Participants will be provided with an Excel tool 5, the Penalties Computation Tool, that will allow them to calculate the penalties using their results from the practice server. 5 The Penalties Computation Tool will be released on Nov 9 th on ROTC website. 11

12 Sample News Release Schedule Time Week Release 1 Week 1 The realized volatility of RTM for this week will be 20% 75 Week 1 The realized volatility of RTM for next week will be between 27-30% 150 Week 2 The realized volatility of RTM for this week will be 29% 450 Week 4 The realized volatility of RTM for this week will be 26% Trading Limits and Transaction Costs Each participant will be subject to gross and net trading limits specific to the security type as specified below. The gross trading limit reflects the sum of the absolute values of the long and short positions across all securities; while the net trading limit reflects the sum of long and short positions such that short positions negate any long positions. Trading limits will be enforced and participants will not be able to exceed them. Volatility Trading Case Case Security Type Gross Limit Net Limit RTM ETF 50,000 Shares 50,000 Shares RTM Options 2,500 Contracts 1,000 Contracts The maximum trade size will be 10,000 shares for RTM and 100 contracts for RTM options, restricting the volume of shares and contracts transacted per trade to 10,000 and 100 respectively. Transaction fees will be set at $0.02 per share traded for RTM and $2.00 per contract traded for RTM options. As with standard options markets, each contract represents 100 shares (purchasing 1 option contract for $0.35 will actually cost $35 plus a $2 commission, and will settle based on the exercise value of 100 shares). Position Close-Out Any outstanding position in RTM will be closed at the end of trading based on the last-traded price. There are no liquidity constraints for either the options or RTM. All options will be cash-settled based on the following: Where, SS is the last price of RTM; KK is the strike price of the option. CCCCCCCC OOOOOOOOOOOO PPPPPPPPPPPP = mmmmmm{0, SS KK} PPPPPP OOOOOOOOOOOO PPPPPPPPPPPP = mmmmmm{0, KK SS} 12

13 Key Objectives Objective 1 Build a model to forecast the future volatility of the underlying ETF based on known information and given forecast ranges. Participants should use this model with an options pricing model to determine whether the market prices for options are overvalued or undervalued. They should then trade the specific options accordingly. Objective 2 Use Greeks to calculate the portfolio exposure and hedge the position to reduce the risk of the portfolio while profiting from volatility differentials across options. Objective 3 Participants should also seek arbitrage opportunities across different options. Volatility Trading Case 13

14 Algorithmic Trading Case OVERVIEW The Algorithmic Trading Case is designed to challenge participants programming skills since they are required to develop algorithms using RIT API to automate the market-making process and react to changing market conditions. Throughout the case, these algorithms will submit orders to profit from market-making and also from any arbitrage opportunities that may arise. Due to the high-frequency nature of the case, participants will have to develop algorithms to adapt to changes in market dynamics. DESCRIPTION There will be five, 5-minute heats with each heat to be independently traded and representing one day of trading. Algorithmic Trading Case Parameter Value Value Number of trading heats 5 Trading time per heat 300 seconds (5 minutes) Calendar time per heat 1 day of trading All trades must be automatically executed by a trading algorithm. Participants will not be allowed to trade manually through the RIT Client. However, they are allowed to and encouraged to use and modify their algorithms in response to prevailing market conditions and competition from the algorithms of other teams. In addition, there will be 2 minutes in between each heat to alter the algorithms. A base template algorithm will be provided for participants on the ROTC website and can be directly modified for use in the competition. Alternatively, participants can create their own algorithm using RIT API. MARKET DYNAMICS This case will involve 3 stocks and 1 ETF with varying levels of volatility and liquidity. This dynamic exposes participants to the basics of market microstructure in the context of algorithmic trading. Participants can manage the market price impact of trades and automate market making operations by dividing larger volume orders into smaller trades and submitting pairs of trades electronically. The ETF pricing will reflect the following weighted sum of the 3 stocks traded, subject to periodic shocks to its price. CND = BEAV + 2 MAPL + 2 COLD 14

15 Participants will be able to trade 4 securities, the details of which are shown below. Ticker BEAV MAPL COLD CND Starting Price $15 $15 $15 $75 Fee/share (Market orders) $0.01 $0.01 $0.01 $0.02 Rebate/share (Limit/Passive $0.005 $0.005 $0.005 $0.015 orders) Max order size 10,000 10,000 10,000 5,000 Annualized volatility 26% 17% 35% 16% Liquidity Medium High Low Medium Type Stock Stock Stock ETF There will be no information provided that will allow participants to predict the future price or direction of any security. A fee-rebate structure will be instituted to compensate participants for the addition of liquidity through limit orders. As such, participants will have the opportunity to generate returns by market making. Algorithmic Trading Case TRADING LIMITS AND TRANSACTION COSTS Time of Sub-heat (tick) 0 ~ Gross/Net 200,000/50, ,000/0 Each participant will be subject to gross and net trading limits which will change over each heat. The gross trading limit reflects the sum of the absolute values of the long and short positions across all securities and the net trading limit reflects the sum of long and short positions such that short positions negate any long positions. Trading limits will be strictly enforced and participants will not be able to exceed them. Each position in stock will be counted towards trading limits with a multiplier of 1, while each position in the ETF will be counted with a multiplier of 5 (i.e. if you long 100 shares of any stocks, your gross and the net trading limits will increase by 100. If you long 100 positions of CND, your gross and net trading limits will increase by 500 (100 positions * multiplier of 5). Participants will be penalized for having a non-zero net position at the end of any individual heat. This penalty will be levied against the final P&L for that heat. The calculation for the penalty is as follows: Penalty = NNNNNN pppppppppppppppp tt=300 * $

16 The maximum trade size will be 10,000 shares for stocks and 5,000 shares for the ETF, restricting the volume of shares transacted per trade to 10,000 shares for stocks and 5,000 shares for the ETF. Transaction fees will be set at $0.01 per share for stocks and $0.02 per share for the ETF on all market orders filled. Subsequently, a rebate of $0.005 per share for stocks and $0.015 per share for the ETF will be given for all submitted limit orders that are filled. Due to this rebate structure, it is possible for a participant to be profitable after buying and selling a security at the same price, provided that the participant uses limit orders. POSITION CLOSE-OUT Any non-zero position in either stock will be closed out at the end of trading based on the last traded price. It is strongly suggested that participants close out their positions prior to the end of trading period as open positions will be subject to the penalty described above. KEY OBJECTIVE Edit the template algorithm provided and optimize the trading parameters such that the algorithm efficiently balances positions while submitting orders to profit from capturing bid-ask spreads and/or from capturing arbitrage opportunities. Consider rewriting and redesigning the algorithm using your own logic. Since the same template is being provided to all participants, there is a limitation on your ability to differentiate yourself by simply modifying the base template, hence incentive to create your own algorithm using RIT API. Algorithmic Trading Case Case 16